US8170368B2 - Correcting device and method for perspective transformed document images - Google Patents
Correcting device and method for perspective transformed document images Download PDFInfo
- Publication number
- US8170368B2 US8170368B2 US12/076,122 US7612208A US8170368B2 US 8170368 B2 US8170368 B2 US 8170368B2 US 7612208 A US7612208 A US 7612208A US 8170368 B2 US8170368 B2 US 8170368B2
- Authority
- US
- United States
- Prior art keywords
- vertical
- vanishing point
- horizontal
- line segment
- unit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related, expires
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/247—Aligning, centring, orientation detection or correction of the image by affine transforms, e.g. correction due to perspective effects; Quadrilaterals, e.g. trapezoids
Definitions
- the present invention relates to a correcting device and a correcting method for perspective transformation of document images, and more particularly, to a hybrid method combining the detection of vanishing points (including horizontal vanishing points and vertical vanishing points) by voting of various line segments and the detection of vanishing points by using image projection analysis, a method for searching vertical vanishing points by clustering based on vertical strokes of characters, and a method for correcting perspective transformation based on text knowledge.
- Perspective transformation correction of document images based on a digital camera is an important step in document analysis and recognition.
- the meaning of a perspective transformation correction operation is to convert images having perspective transformation into images without perspective transformation.
- the present invention is made in view of the aforementioned defects and limiting of the prior art.
- the present invention proposes a correcting device and a correcting method for perspective transformations based on text knowledge.
- a correcting device for a perspective transformed document image.
- the correcting device comprises a horizontal vanishing point determining unit, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining unit, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting unit, for correcting the perspective transformed document image; wherein the horizontal vanishing point determining unit comprises a direct horizontal line segment detecting unit, an indirect horizontal line segment detecting unit and a horizontal vanishing point detecting unit, and wherein the horizontal vanishing point detecting unit detects a horizontal vanishing point in accordance with a direct horizontal line segment detected by the direct horizontal line segment detecting unit and an indirect horizontal line segment detected by the indirect horizontal line segment detecting unit.
- a correcting device for a perspective transformed document image.
- the correcting device comprises a horizontal vanishing point determining unit, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining unit, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting unit, for correcting the perspective transformed document image; wherein the vertical vanishing point determining unit comprises a direct vertical line segment detecting unit, an indirect vertical line segment detecting unit and a vertical vanishing point detecting unit, and wherein the vertical vanishing point detecting unit detects a vertical vanishing point in accordance with a direct vertical line segment detected by the direct vertical line segment detecting unit and an indirect vertical line segment detected by the indirect vertical line segment detecting unit.
- the correcting method comprises a horizontal vanishing point determining step, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining step, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting step, for correcting the perspective transformed document image; wherein the horizontal vanishing point determining step comprises a direct horizontal line segment detecting step, an indirect horizontal line segment detecting step and a horizontal vanishing point detecting step, and wherein the horizontal vanishing point detecting step detects a horizontal vanishing point in accordance with a direct horizontal line segment detected by the direct horizontal line segment detecting step and an indirect horizontal line segment detected by the indirect horizontal line segment detecting step.
- the correcting method comprises a horizontal vanishing point determining step, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining step, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting step, for correcting the perspective transformed document image; wherein the vertical vanishing point determining step comprises a direct vertical line segment detecting step, an indirect vertical line segment detecting step and a vertical vanishing point detecting step, and wherein the vertical vanishing point detecting step detects a vertical vanishing point in accordance with a direct vertical line segment detected by the direct vertical line segment detecting step and an indirect vertical line segment detected by the indirect vertical line segment detecting step.
- the methods according to the present invention remove the problem of high computational complexity of the direct methods, and overcome at the same time the defect of instability of the indirect methods.
- the correcting method for perspective transformed documents based on a synthesized vanishing points detecting method is a fast and robust correcting method for a perspective transformed document image.
- FIG. 1 is a structural diagram showing the correcting device for a perspective transformed document image according to this invention
- FIG. 2 is a flowchart showing the correcting method for a perspective transformed document image according to this invention
- FIG. 3 shows a simple and quick Smearing algorithm
- FIG. 4 is a structural diagram showing a horizontal text line detecting unit
- FIG. 5 is a flowchart showing a horizontal text line detecting process
- FIG. 6 is a diagram showing a horizontal text line detecting device based on shape and size analysis of connected components
- FIG. 7 is a flowchart showing horizontal text line detection based on shape and size analysis of connected components
- FIG. 8 is a structural diagram showing the horizontal vanishing point detecting unit
- FIG. 9 is a flowchart showing horizontal vanishing point detection
- FIG. 10 is a certain perspective transformed document image
- FIG. 11 is a diagram showing various horizontal line segments of the image in FIG. 10 ;
- FIG. 12 is a diagram showing distribution of intersections in pairs of the horizontal line segments in FIG. 11 ;
- FIG. 13 is a structural diagram showing a character vertical stroke detecting unit
- FIG. 14 is a flowchart showing character vertical stroke detection
- FIG. 15 is a structural diagram showing the vertical vanishing point detecting unit.
- FIG. 16 is a flowchart showing vertical vanishing point detection.
- FIG. 1 is a structural diagram showing the correcting device for a perspective transformed document image according to one embodiment of this invention.
- the correcting device for a perspective transformed document image according to this invention comprises a perspective transformed document image inputting unit 101 , an image diminishing unit 102 , a border detecting and binarizing unit 103 , a direct horizontal and vertical line segment detecting unit 104 (corresponding to the direct horizontal line segment detecting unit and the direct vertical line segment detecting unit according to this invention), a horizontal text line detecting unit 105 , a horizontal vanishing point detecting unit 106 , a character vertical stroke detecting unit 107 , a vertical vanishing point detecting unit 108 , a perspective transformation correcting and converting unit 109 , and a corrected image outputting unit 110 .
- FIG. 2 is a flowchart showing the correcting method for a perspective transformed document image according to this invention.
- a perspective transformed document image is inputted by the perspective transformed document image inputting unit 101 (step 201 ) in the correcting device for a perspective transformed document image according to an embodiment of this invention.
- the document image can be derived from a digital camera, a mobile phone equipped with a pickup lens, or other photographing devices.
- the inputted document image is inputted into the image diminishing unit 102 , which performs grayscale and diminishing operations on the image (step 202 ) to enhance the processing speed of the system.
- the image is inputted into the border detecting and binarizing unit 103 .
- the image can be directly inputted into the border detecting and binarizing unit 103 without passing through the image diminishing unit 102 .
- the border detecting and binarizing unit 103 calculates to obtain the border image of the grayscale image, and converts the grayscale image into a binarized image (steps 203 and 211 ).
- the border detecting and binarizing unit 103 may employ various conventional image bordering algorithms and binarizing algorithms, such as the effective Canny bordering algorithm and the high-speed Otsu binarizing algorithm.
- the document images and perspective transformed document images in the present invention may be document images inputted externally (for instance via a digital camera, a mobile phone equipped with a pickup lens or other photographing devices), or documents images having been processed by the border detecting and binarizing unit according to the context.
- the direct horizontal and vertical line segment detecting unit 104 detects to obtain a long horizontal line segment (referred to as a direct horizontal line segment 207 ) and a long vertical line segment (referred to as a direct vertical line segment 206 ) by performing connected components analysis on the border image.
- a long horizontal line segment referred to as a direct horizontal line segment 207
- a long vertical line segment referred to as a direct vertical line segment 206
- the direct horizontal and vertical line segment detecting unit 104 if the long axis direction of a connected component is adjacent to the horizontal direction, the length of the long axis is greater than a threshold value and the length of the short axis is less than another threshold value, then the long axis of this connected component is a long horizontal line segment.
- the long axis direction of a connected component is adjacent to the vertical direction, the length of the long axis is greater than a threshold value and the length of the short axis is less than another threshold value, then the long axis of this connected component is a long vertical line segment.
- detections of the direct horizontal line segment and the direct vertical line segment are both performed by the direct horizontal and vertical line segment detecting unit 104 , there can also be two individual units, one performing detection of the direct horizontal line segment and the another performing detection of the direct vertical line segment.
- the horizontal text line detecting unit 105 detects the horizontal text lines in the border image, and obtains a corresponding horizontal line segment (referred to as an indirect horizontal line segment 208 ) from these text lines.
- the horizontal text line detecting unit 105 detects the horizontal text lines by means of a simple and quick Smearing algorithm, for example, as shown in FIG. 3 .
- the horizontal vanishing point detecting unit 106 detects the vanishing points and obtains the final horizontal vanishing point 210 on the basis of the direct horizontal line segment 207 and the indirect horizontal line segment 208 and in combination with a direct method based on image projection analysis and an indirect method based on the analysis and voting of the horizontal line segments and the intersection thereof. This will be described in greater detail in the following paragraphs.
- step 213 the character vertical stroke detecting unit 107 performs character vertical stroke detection on the border image obtained by the border detecting and binarizing unit 103 in step 211 , to obtain an indirect vertical line segment 214 .
- the unit 107 is employed to obtain the indirect vertical line segment 214 , while in other embodiments of the invention, this unit can employ other methods such as those disclosed in the prior art. Even under such a circumstance, since the present invention makes use of the synthesized method that combines the direct method and the indirect method, the object of the invention can still be achieved.
- step 211 differs from step 203 in the fact that the image in step 211 has not been diminished.
- the transformed document image 212 is identical with the transformed document image 201 .
- the character vertical stroke detecting unit 107 obtains the vertical strokes of the characters by performing connected components analysis on the border image, so as to obtain a line segment having vertical direction indication (namely the indirect vertical line segment 214 ).
- the vertical vanishing point detecting unit 108 detects vanishing points and obtains the final vertical vanishing point (namely the vertical vanishing point 216 ) on the basis of the direct vertical line segment 206 and the indirect vertical line segment 214 and in combination with a direct method based on image projection analysis and an indirect method based on the analysis and voting of the vertical line segments and the intersection thereof.
- the direct horizontal and vertical line segment detecting unit 104 (which detects a portion of the direct horizontal line segments), the horizontal text line detecting unit 105 and the horizontal vanishing point detecting unit 106 correspond to the horizontal vanishing point determining unit, while the direct horizontal and vertical line segment detecting unit 104 (which detects a portion of the direct vertical line segments), the character vertical stroke detecting unit 107 and the vertical vanishing point detecting unit 108 correspond to the vertical vanishing point determining unit.
- the perspective transformation correcting and converting unit 109 obtains a conversion matrix of the perspective transformation by means of the horizontal vanishing point and the vertical vanishing point, and obtains the final corrected document image (namely a transformation corrected image 218 ) by performing a correction and conversion processing based on text knowledge, and the final corrected document image is outputted by the outputting unit 110 .
- FIG. 3 shows a simple and quick Smearing algorithm (represented by C Language) used by the horizontal text line detecting unit 105 in step 205 as shown in FIG. 2 .
- ‘height’ represents the height of a banarized image to be processed
- ‘width’ represents the width thereof
- This quick algorithm horizontally scans each line of images. In each line, if the distance between two adjacent black pixel points is less than a certain threshold value (smear_thres), the pixel points between these two points are all assigned to be black pixel points.
- FIG. 4 is a structural diagram showing the horizontal text line detecting unit 105 .
- the horizontal text line detecting unit 105 includes a binarized document image inputting unit 401 , an image Smearing processing unit 402 , a connected component calculating unit 403 , a horizontal text line detecting unit 404 based on analysis of the shape and size of the connected component, and an indirect horizontal line segment outputting unit 405
- FIG. 5 is an operational flowchart of the horizontal text line detecting unit 105 .
- a binarized image having been performed with border detection and binarization by the border detecting and binarizing unit 103 is inputted by the inputting unit 401 .
- the image Smearing processing unit 402 performs Smearing processing by means of the simple and quick Smearing algorithm, for example, as shown in FIG. 3 .
- the Smearing algorithm scans, in the horizontal direction, the image, and analyzes the relationship between the black point pixels in the horizontal direction: if the distance between two black point pixels in the horizontal direction is less than a predetermined threshold value, the pixels between these two points all become black point pixels.
- the connected component calculating unit 403 calculates the connected component processed by the Smearing by analysing the interrelationship between the black point pixels.
- the horizontal text line detecting unit 404 based on analysis of the shape and size of the connected component detects the horizontal text lines by analyzing the size, shape and direction of the connected components, and the specific procedures thereof are as shown in FIG. 7 .
- the indirect horizontal line segment outputting unit 405 outputs the obtained horizontal line segment (namely the indirect horizontal line segment 208 ) which represents the horizontal text lines.
- FIG. 6 is a structural diagram of the horizontal text line detecting unit 404 based on analysis of the shape and size of the connected component.
- the horizontal text line detecting unit 404 based on analysis of the shape and size of the connected component includes a long connected component selecting unit 601 , a connected component baseline calculating unit 602 , and a baseline analyzing unit 603 .
- FIG. 7 is an operational flowchart of the horizontal text line detecting unit 404 based on analysis of the shape and size of the connected component.
- the unit 601 selects a long connected component, for instance, selects, as a candidate horizontal text line, a connected component having a relatively long length (namely the length being greater than a certain threshold value).
- the unit 602 calculates the upper, middle and lower baselines of the connected component, namely calculates the upper, middle and lower baselines with regard to the connected component of the candidate text line.
- the specific step is as follows: the upper and lower contour points of the connected component are first calculated, and the sequence is ⁇ (x 1 , y 1 U ), (x 2 , y 2 U ), . . . , (x N , y N U ) ⁇ , ⁇ (x 1 , y 1 L ), (x 2 , y 1 L ), . . . , (x N , y N L ) ⁇ , where (x,y) represents the image coordinates, and N is the length of this connected component.
- the unit 703 analyzes the directional relationship between the upper baseline and the lower baseline: if the upper and lower baselines of a connected component are substantially in the same direction (namely the angle therebetween being less than a predetermined threshold value), and a difference between the average height, to which all contour points in the upper and lower baselines correspond, and the height of a standard text line is less than a lesser threshold value of a certain value, it is indicated that this connected component is a horizontal text line.
- a line segment obtained by fitting the sequences of the contour points can be used as the direct horizontal line segment (namely the direct horizontal line segment 207 ) represented by this text line.
- FIG. 8 is a structural diagram of the horizontal vanishing point detecting unit 106 .
- This unit is a synthesized device combining the direct method based on image projection analysis and the indirect method of voting of various horizontal line segments.
- the horizontal vanishing point detecting unit 106 includes a horizontal line segment inputting unit 801 , a horizontal line segment intersection clustering unit 802 , a candidate horizontal vanishing point selecting unit 803 , a horizontal direction perspective projection analyzing unit 804 , a horizontal vanishing point synthesis analyzing unit 805 , and a horizontal vanishing point outputting unit 806 .
- FIG. 9 is an operational flowchart of the horizontal vanishing point detecting unit 106 .
- the method is a synthesized method combining the direct method based on image projection analysis and the indirect method of voting of various horizontal line segments.
- the inputting unit 801 inputs the direct horizontal line segment 207 and the indirect horizontal line segment 208 .
- horizontal vanishing point detection is performed by using an indirect method similar to the voting of line segments and the intersection thereof.
- the horizontal line segment intersection clustering unit 802 clusters the point collections formed by intersections in pairs of all horizontal line segments (including the obtained direct horizontal line segment and indirect horizontal line segment) by a clustering method (such as the K-Means method) to obtain a plurality of clusters. All these intersections have such a property that they can be better clustered into a plurality of sub-collections as shown in FIGS. 10-12 .
- FIG. 10 shows an original perspective transformed document image
- FIG. 11 shows the horizontal line segments detected and obtained by a horizontal text line detecting unit 205
- step 903 the candidate horizontal vanishing point selecting unit 803 selects the central point of each cluster as a candidate horizontal vanishing point, and takes a ratio of the number of the intersections contained in this cluster to the number of all intersections as a weighting coefficient of the cluster of the candidate horizontal vanishing point.
- the coefficient is set as f h d (k), where k represents the k th candidate horizontal vanishing point.
- a direct method similar to image projection analysis is employed to perform horizontal vanishing point detection on the aforementioned collections of the candidate horizontal vanishing points.
- a projection method is also employed in this invention to perform analysis on the candidate horizontal vanishing points. That is to say, in step 904 the horizontal direction perspective projection analyzing unit 804 performs perspective projection analysis in the horizontal direction on the document image with regard to the horizontal vanishing points.
- the projection method employed by the system on each of the candidate horizontal vanishing points selected by the candidate horizontal vanishing point selecting unit 803 is the same as the method presented in the article “Rectifying perspective views of text in 3D scenes using vanishing points” by P. Clark, and M. Mirmehdi in section 3 of Pattern Recognition 36(11), 2003.
- the derivative-squared-sum of the projection value of each candidate horizontal vanishing point is obtained, and a ratio of the derivative-squared-sum of each candidate horizontal vanishing point to the derivative-squared-sum of all candidate points is taken as a weighting coefficient of the projection analysis of this candidate point.
- the coefficient is set as f h i (k), where k represents the k th candidate horizontal vanishing point.
- the candidate horizontal vanishing point having the greatest synthesized weighting coefficient is selected as the final horizontal vanishing point, and outputted by the outputting unit 806 (step 906 ).
- This method removes the problem of high computational complexity of the direct methods, and overcomes at the same time the defect of instable performance of the indirect methods.
- This synthesized method is a fast and robust method for detecting the vanishing points.
- FIG. 13 is a structural diagram showing the character vertical stroke detecting unit 107 according to one embodiment of the present invention.
- the character vertical stroke detecting unit 107 includes a vertical border image inputting unit 1301 , a border image connected component calculating unit 1302 , a vertical stroke detecting unit 1303 , and an indirect vertical line segment outputting unit 1304 .
- FIG. 14 is an operational flowchart of the character vertical stroke detecting unit 107 according to one embodiment of the present invention.
- the inputting unit 1301 inputs a vertical border image obtained by the border detecting and binarizing unit 103 .
- the border image connected component calculating unit 1302 calculates the connected components with regard to the vertical border image.
- the vertical stroke detecting unit 1303 analyzes the shapes and sizes of the connected components, selects a connected component having a height similar to the height of the characters and having a direction adjacent to the vertical direction as the candidate vertical stroke, and analyzes the shape of this candidate vertical stroke connected component.
- the height of the connected component when the absolute value of the difference between the height of the connected component and the height of the characters is less than a predetermined threshold value, this is referred to as the height of the connected component being similar to the height of the characters.
- the absolute value of the difference between the direction of the connected component and the vertical direction is less than a predetermined threshold value, this is referred to as the direction of the connected component being adjacent to the vertical direction.
- DIS i ⁇ ( x , y ) ⁇ a i ⁇ y + b i ⁇ x + c a 2 + b 2 ⁇
- f ⁇ ( LC i ) ⁇ 1 ⁇ N LC i > n_thres ⁇ _stroke 0 ⁇ otherwise ⁇ ⁇
- N(x, ⁇ , ⁇ ) is a Gaussian distribution of the line segment LC, with the means being ⁇ , and the standard variance being ⁇ stroke and ⁇ stroke are the average value and the standard variance relevant to the character vertical stroke obtained empirically and experimentally.
- p_thres_stroke is a threshold value approximately equal to 1, and can be set as 0.98.
- n_thres_stroke is approximately equal to the number of the black point pixels in this connected component. If f(LC i ) 1 , it is indicated that C i is the character vertical stroke. By this time, these line segments fitted by the connected component of the character vertical stroke are the indirect vertical line segments.
- the outputting unit 1304 performs the output.
- FIG. 15 is a structural diagram showing the vertical vanishing point detecting unit 108 according to one embodiment of this invention.
- This unit is a synthesized device combining the direct method based on image projection analysis and the indirect method of voting of various vertical line segments.
- the vertical vanishing point detecting unit 108 according to one embodiment of this invention includes a vertical line segment inputting unit 1501 , a vertical line segment intersection clustering unit 1502 , a candidate vertical vanishing point selecting unit 1503 , a vertical direction perspective projection analyzing unit 1504 , a vertical vanishing point synthesis analyzing unit 1505 , and a vertical vanishing point outputting unit 1506 .
- FIG. 16 is a flowchart of the vertical vanishing point detecting unit 108 according to one embodiment of this invention.
- the method is a synthesized method combining the direct method based on image projection analysis and the indirect method of voting of various vertical line segments.
- the inputting unit 1501 inputs the direct vertical line segment 206 and the indirect vertical line segment 214 .
- vertical vanishing point detection is performed by using an indirect method similar to the voting of the line segments and the intersection thereof.
- the vertical line segment intersection clustering unit 1502 clusters the point collections formed by intersections in pairs of all vertical line segments (including the obtained direct vertical line segment and indirect vertical line segment) by a clustering method (such as the K-Means method) to obtain a plurality of clusters. All these intersections have such a property that they can be better clustered into a plurality of sub-collections.
- the candidate vertical vanishing point selecting unit 1503 selects the central point of each cluster as a candidate vertical vanishing point, and takes a ratio of the number of the intersections contained in this cluster to the number of all intersections as a weighting coefficient of the cluster of the candidate vertical vanishing point.
- the coefficient is set as f v d (k), where k represents the k th candidate vertical vanishing point.
- a direct method similar to image projection analysis is employed to perform candidate vanishing point detection on the aforementioned collections of the candidate vertical vanishing points.
- the vertical direction perspective projection analyzing unit 1504 performs perspective projection analysis in the vertical direction.
- the projection method employed by the system on each of the candidate vertical vanishing points obtained by the selecting unit 1503 is similar to the method presented in the article “Rectifying perspective views of text in 3D scenes using vanishing points” by P. Clark, and M. Mirmehdi in section 3 of Pattern Recognition 36(11), 2003.
- the projection is not directed to the whole image, but directed to each line of the horizontal text lines (the horizontal text lines here indicate the documents row-by-row in the document image, and can be obtained by the previously mentioned horizontal document line detecting unit).
- N is the number of the text lines
- l is the number of bin
- the text lines here are obtained by the horizontal text line detecting unit 105 .
- the candidate vertical vanishing point having the greatest synthesized weighting coefficient is selected as the final vertical vanishing point, and the vertical vanishing point is outputted by the outputting unit 1506 in step 1606 .
- Text regions are obtained through text detection, and correction and conversion are performed on these text regions based on interpolation (such as linear interpolation) while other regions are replaced directly with adjacent points.
- interpolation such as linear interpolation
- the text region here is obtained by the horizontal text line detecting unit 205 .
- the result outputted by the system is an image whose transformation has been corrected.
- the present invention proposes a method that searches the vertical vanishing points based on clustering the vertical strokes of characters, detects the vertical strokes of characters through connected component analysis of the shapes and sizes of the character strokes, and clusters the intersections of all the vertical strokes in pairs to obtain a plurality of clusters, the center of each of which is a candidate vertical vanishing point.
- This method has stronger robustness than the method utilizing analysis of the vertical distance between horizontal line segments or vertical line segments to detect the vertical vanishing points.
- the perspective transformation correcting and converting method based on text knowledge as proposed in the present invention performs conversion based on interpolation processing merely on the region having text of the transformed image. This method enhances the speed of the whole device and the method.
- the vanishing point detection in the method uses a synthesized method that combines the direct method based on image projection analysis and the indirect method for detecting vanishing points by voting of various line segments, makes use of various segments to perform voting and clustering to obtain a plurality of candidate points of the vanishing points, performs image projection analysis on these candidate points, combines the results of the foregoing two steps to obtain the final vanishing points (including the horizontal vanishing points and the vertical vanishing points), and performs corresponding perspective transformation correction.
- the method clusters the vertical strokes of characters to search the vertical vanishing points, detects reliable vertical strokes of characters by using rule-based connected component analysis, and clusters the intersections of these vertical strokes to obtain a plurality of candidate points of the vertical vanishing points.
- the method performs perspective transformation correction and conversion based on text knowledge, obtains a transformation correction and conversion matrix from the horizontal vanishing points and the vertical vanishing points, merely converts the region having text of the transformed image, and replaces other regions with adjacent points.
- the present invention is applicable for document images shot by a digital camera, by a mobile phone with a pickup lens, and by other photographing devices.
- a correcting device for a perspective transformed document image.
- the correcting device comprises a horizontal vanishing point determining unit, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining unit, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting unit, for correcting the perspective transformed document image; wherein the horizontal vanishing point determining unit comprises a direct horizontal line segment detecting unit, an indirect horizontal line segment detecting unit and a horizontal vanishing point detecting unit, and wherein the horizontal vanishing point detecting unit detects a horizontal vanishing point in accordance with a direct horizontal line segment detected by the direct horizontal line segment detecting unit and an indirect horizontal line segment detected by the indirect horizontal line segment detecting unit.
- a correcting device for a perspective transformed document image.
- the correcting device comprises a horizontal vanishing point determining unit, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining unit, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting unit, for correcting the perspective transformed document image; wherein the vertical vanishing point determining unit comprises a direct vertical line segment detecting unit, an indirect vertical line segment detecting unit and a vertical vanishing point detecting unit, and wherein the vertical vanishing point detecting unit detects a vertical vanishing point in accordance with a direct vertical line segment detected by the direct vertical line segment detecting unit and an indirect vertical line segment detected by the indirect vertical line segment detecting unit.
- the correcting method comprises a horizontal vanishing point determining step, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining step, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting step, for correcting the perspective transformed document image; wherein the horizontal vanishing point determining step comprises a direct horizontal line segment detecting step, an indirect horizontal line segment detecting step and a horizontal vanishing point detecting step, and wherein the horizontal vanishing point detecting step detects a horizontal vanishing point in accordance with a direct horizontal line segment detected by the direct horizontal line segment detecting step and an indirect horizontal line segment detected by the indirect horizontal line segment detecting step.
- the correcting method comprises a horizontal vanishing point determining step, for detecting a horizontal vanishing point of the perspective transformed document image; a vertical vanishing point determining step, for detecting a vertical vanishing point of the perspective transformed document image; and a perspective transformation correcting and converting step, for correcting the perspective transformed document image; wherein the vertical vanishing point determining step comprises a direct vertical line segment detecting step, an indirect vertical line segment detecting step and a vertical vanishing point detecting step, and wherein the vertical vanishing point detecting step detects a vertical vanishing point in accordance with a direct vertical line segment detected by the direct vertical line segment detecting step and an indirect vertical line segment detected by the indirect vertical line segment detecting step.
- the horizontal line segment intersection clustering unit makes use of the K-Means method to perform the clustering.
- the horizontal vanishing point detecting unit includes a horizontal line segment intersection clustering unit that clusters the point collections formed by intersections in pairs of all horizontal line segments to obtain a plurality of clusters; a candidate horizontal vanishing point selecting unit that selects the central point of each cluster as a candidate horizontal vanishing point, and obtains a weighting coefficient of each candidate horizontal vanishing point; a horizontal direction perspective projection analyzing unit that performs perspective projection analysis in the horizontal direction on the document image with regard to the candidate horizontal vanishing points, and obtains another weighting coefficient of each candidate horizontal vanishing point; and a horizontal vanishing point synthesis analyzing unit that selects the final horizontal vanishing point based on the weighting coefficient and the another weighting coefficient.
- the horizontal vanishing point synthesis analyzing unit synthesizes the weighting coefficient obtained by the horizontal direction perspective projection analyzing unit and the weighting coefficient obtained by the candidate horizontal vanishing point selecting unit in a linear mode.
- the vertical vanishing point determining unit comprises a direct vertical line segment detecting unit, an indirect vertical line segment detecting unit and a vertical vanishing point detecting unit, and wherein the vertical vanishing point detecting unit detects a vertical vanishing point in accordance with a direct vertical line segment detected by the direct vertical line segment detecting unit and an indirect vertical line segment detected by the indirect vertical line segment detecting unit.
- the indirect vertical line segment detecting unit includes a connected component calculating unit that calculates a connected component of the document image; and a character vertical stroke detecting unit that analyzes the shape and size of the connected component to determine a candidate vertical stroke, and analyzes the shape of the connected component of the candidate vertical stroke to obtain a vertical line segment.
- the character vertical stroke detecting unit selects a connected component having a height similar to the height of the characters and having a direction adjacent to the vertical direction as the candidate vertical stroke.
- the vertical vanishing point detecting unit includes a vertical line segment intersection clustering unit that clusters the point collections formed by intersections of vertical line segments in pairs from the collection composed by direct vertical line segments and indirect vertical line segments to obtain a plurality of clusters; a vertical vanishing point selecting unit that selects the central point of each cluster as a candidate vertical vanishing point, and obtains a weighting coefficient of each candidate vertical vanishing point; a vertical direction perspective projection analyzing unit that performs perspective projection analysis in the vertical direction on the document image with regard to the candidate vertical vanishing points, and obtains another weighting coefficient of each candidate vertical vanishing point; and a vertical vanishing point synthesis analyzing unit that obtains the final vertical vanishing point by analyzing the weighting coefficient obtained by the vertical vanishing point selecting unit and the another weighting coefficient obtained by the vertical direction perspective projection analyzing unit with regard to each candidate vertical vanishing point.
- the vertical line segment intersection clustering unit makes use of the K-Means method to perform the clustering.
- the vertical direction perspective projection analyzing unit performs the projection on each line of the horizontal text line, and synthesizes the results of the projecting analysis of all text lines.
- the vertical vanishing point selecting unit takes a ratio of the number of the intersections contained in the cluster to the number of all intersections as the weighting coefficient of the cluster of the candidate vertical vanishing point.
- the present invention further provides a computer program which is executable by a computer to carry out the correcting method of the perspective transformed document image according to this invention.
- the present invention further provides a computer program which is executable by a computer to make the computer function as the correcting device of the perspective transformed document image according to this invention.
- the present invention provides a data storage medium which stores thereon said computer program.
- the storage medium can be any storage media known to a person skilled in the art, such as ROM, floppy disk, flash memory, hard disk, CD, DVD, tape, etc.
- FIG. 1 101 —perspective transformed document image inputting unit; 102 —image diminishing unit; 103 —border detecting and binarizing unit; 104 —direct horizontal and vertical line segments detecting unit; 105 —horizontal text line detecting unit; 106 —horizontal vanishing point detecting unit; 107 —character vertical stroke detecting unit; 108 —vertical vanishing point detecting unit; 109 —perspective transformation correcting and converting unit; 110 —corrected image outputting unit
- FIG. 2 201 —perspective transformed document image; 202 —image diminishing; 203 —border detection and binarization; 204 —direct horizontal and vertical line segments detection; 205 —horizontal text line detection; 206 —direct vertical line segment; 207 —direct horizontal line segment, 208 —indirect horizontal line segment, 209 —horizontal vanishing point detection; 210 —horizontal vanishing point, 211 —border detection and binarization again; 212 —original transformed document image; 213 —character vertical stroke detection; 214 —indirect vertical line segment; 215 —vertical vanishing point detection; 216 —vertical vanishing point, 217 —perspective transformation correction and conversion; 218 —transformation corrected image
- FIG. 7 701 —selection of long connected component; 702 —calculation of upper, middle and lower baselines of connected component; 703 —analysis of directional relationship between upper and lower baselines
- FIG. 8 801 —horizontal line segment inputting unit; 802 —horizontal line segment intersection clustering unit; 803 —candidate horizontal vanishing point selecting unit, 804 —horizontal direction perspective projection analyzing unit, 805 —horizontal vanishing point synthesis analyzing unit, 806 —horizontal vanishing point outputting unit
- FIG. 9 901 —direct and indirect horizontal line segments; 902 —horizontal line segment intersection clustering; 903 —candidate horizontal vanishing point selection; 904 —horizontal direction perspective projection analysis; 905 —horizontal vanishing point synthesis analysis; 906 —horizontal vanishing point
- FIG. 13 1301 —vertical border image inputting unit, 1302 —border image connected component calculating unit; 1303 —character vertical stroke detecting unit; 1304 —indirect vertical line segment outputting unit
- FIG. 15 1501 —vertical line segment inputting unit, 1502 —vertical line segment intersection clustering unit, 1503 —candidate vertical vanishing point selecting unit; 1504 —vertical direction perspective projection analyzing unit, 1505 —vertical vanishing point synthesis analyzing unit; 1506 —vertical vanishing point outputting unit
- FIG. 16 1601 —direct and indirect vertical line segments; 1602 —vertical line segment intersection clustering; 1603 —candidate vertical vanishing point selection; 1604 —vertical direction perspective projection analysis; 1605 —vertical vanishing point synthesis analysis; 1606 —vertical vanishing point
Abstract
Description
f h(k)=G(f h d(k),f h i(k))
f h(k)=αf h d(k)+βf h i(k)
α+β=1
f v(k)=G(f v d(k),f v i(k))
f v(k)=αf v d(k)+βf v i(k)
α+β=1
i0=int(fi)
i0=int(fj)
image — dst(ij)=image — src(i0j0)
Claims (18)
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200710088355.0 | 2007-03-16 | ||
CN2007100883550A CN101267493B (en) | 2007-03-16 | 2007-03-16 | Correction device and method for perspective distortion document image |
CN200710088355 | 2007-03-16 |
Publications (2)
Publication Number | Publication Date |
---|---|
US20080226171A1 US20080226171A1 (en) | 2008-09-18 |
US8170368B2 true US8170368B2 (en) | 2012-05-01 |
Family
ID=39762758
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/076,122 Expired - Fee Related US8170368B2 (en) | 2007-03-16 | 2008-03-13 | Correcting device and method for perspective transformed document images |
Country Status (3)
Country | Link |
---|---|
US (1) | US8170368B2 (en) |
JP (1) | JP4952625B2 (en) |
CN (1) | CN101267493B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100014782A1 (en) * | 2008-07-15 | 2010-01-21 | Nuance Communications, Inc. | Automatic Correction of Digital Image Distortion |
US20120219221A1 (en) * | 2011-02-25 | 2012-08-30 | Sony Corporation | System and method for effectively performing a scene rectification procedure |
US20120321216A1 (en) * | 2008-04-03 | 2012-12-20 | Abbyy Software Ltd. | Straightening Out Distorted Perspective on Images |
US8897600B1 (en) * | 2013-12-20 | 2014-11-25 | I.R.I.S. | Method and system for determining vanishing point candidates for projective correction |
US20170208207A1 (en) * | 2016-01-20 | 2017-07-20 | Fujitsu Limited | Method and device for correcting document image captured by image pick-up device |
US10140691B2 (en) | 2016-04-26 | 2018-11-27 | Abbyy Development Llc | Correcting perspective distortion in double-page spread images |
US10970845B2 (en) | 2017-06-14 | 2021-04-06 | Canon Kabushiki Kaisha | Image processing apparatus, image processing method, and storage medium |
Families Citing this family (35)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101789122B (en) * | 2009-01-22 | 2013-06-26 | 佳能株式会社 | Method and system for correcting distorted document image |
JP4630936B1 (en) | 2009-10-28 | 2011-02-09 | シャープ株式会社 | Image processing apparatus, image processing method, image processing program, and recording medium recording image processing program |
CN101697235B (en) * | 2009-10-30 | 2013-04-10 | 青岛海信移动通信技术股份有限公司 | Perspective drawing generation method and perspective drawing generation device |
US8947736B2 (en) * | 2010-11-15 | 2015-02-03 | Konica Minolta Laboratory U.S.A., Inc. | Method for binarizing scanned document images containing gray or light colored text printed with halftone pattern |
CN102073997B (en) * | 2011-01-12 | 2012-11-07 | 东华理工大学 | Small-square document image perspective-recovery method |
US9319556B2 (en) | 2011-08-31 | 2016-04-19 | Konica Minolta Laboratory U.S.A., Inc. | Method and apparatus for authenticating printed documents that contains both dark and halftone text |
CN102496018B (en) * | 2011-12-08 | 2013-07-24 | 方正国际软件有限公司 | Document skew detection method and system |
CN103177418B (en) * | 2011-12-22 | 2016-03-02 | 北大方正集团有限公司 | A kind of perspective distortion method and system of planar target image |
CN104657730B (en) * | 2013-11-20 | 2018-01-05 | 富士通株式会社 | Means for correcting, method and the scanner of file and picture |
US8913836B1 (en) * | 2013-12-20 | 2014-12-16 | I.R.I.S. | Method and system for correcting projective distortions using eigenpoints |
JP6542230B2 (en) * | 2013-12-20 | 2019-07-10 | イ.エル.イ.エス. | Method and system for correcting projected distortion |
US8811751B1 (en) * | 2013-12-20 | 2014-08-19 | I.R.I.S. | Method and system for correcting projective distortions with elimination steps on multiple levels |
AU2013273778A1 (en) * | 2013-12-20 | 2015-07-09 | Canon Kabushiki Kaisha | Text line fragments for text line analysis |
JP6010870B2 (en) * | 2013-12-24 | 2016-10-19 | カシオ計算機株式会社 | Image correction apparatus, image correction method, and program |
US9747499B2 (en) * | 2015-03-03 | 2017-08-29 | Fuji Xerox Co., Ltd. | Systems and methods for detection and high-quality capture of documents on a cluttered tabletop with an automatically controlled camera |
CN104835119A (en) * | 2015-04-23 | 2015-08-12 | 天津大学 | Method for positioning base line of bending book cover |
CN106296745B (en) * | 2015-05-26 | 2019-03-12 | 富士通株式会社 | To the corrected method and apparatus of file and picture |
JP6468463B2 (en) * | 2015-07-30 | 2019-02-13 | 京セラドキュメントソリューションズ株式会社 | Image processing device |
CN106803269B (en) * | 2015-11-25 | 2020-03-10 | 富士通株式会社 | Method and device for perspective correction of document image |
CN106910196B (en) * | 2015-12-23 | 2021-01-29 | 北京奇虎科技有限公司 | Image detection method and device |
EP3506167B1 (en) * | 2016-09-12 | 2022-06-22 | Huawei Technologies Co., Ltd. | Processing method and mobile device |
CN107845068B (en) * | 2016-09-18 | 2021-05-11 | 富士通株式会社 | Image view angle conversion device and method |
CN106778739B (en) * | 2016-12-02 | 2019-06-14 | 中国人民解放军国防科学技术大学 | A kind of curving transmogrified text page-images antidote |
US11074679B2 (en) | 2017-02-06 | 2021-07-27 | Huawei Technologies Co., Ltd. | Image correction and display method and device |
US11134170B2 (en) | 2017-12-15 | 2021-09-28 | Hewlett-Packard Development Company, L.P. | Correction of feed skewed images |
RU2680765C1 (en) * | 2017-12-22 | 2019-02-26 | Общество с ограниченной ответственностью "Аби Продакшн" | Automated determination and cutting of non-singular contour of a picture on an image |
CN108492284B (en) * | 2018-03-12 | 2020-03-03 | 百度在线网络技术(北京)有限公司 | Method and apparatus for determining perspective shape of image |
CN109671017A (en) * | 2018-12-25 | 2019-04-23 | 广州励丰文化科技股份有限公司 | A kind of shaped face projection correction method and system based on model |
JP6781996B1 (en) * | 2019-06-12 | 2020-11-11 | 株式会社マーケットヴィジョン | Image correction processing system |
CN110852958B (en) * | 2019-10-11 | 2022-12-16 | 北京迈格威科技有限公司 | Self-adaptive correction method and device based on object inclination angle |
JP7041420B2 (en) * | 2020-09-25 | 2022-03-24 | 株式会社マーケットヴィジョン | Image correction processing system |
CN112634628B (en) * | 2020-12-08 | 2021-10-22 | 鹏城实验室 | Vehicle speed determination method, terminal and storage medium |
CN113096051B (en) * | 2021-04-30 | 2023-08-15 | 上海零眸智能科技有限公司 | Map correction method based on vanishing point detection |
CN113487542B (en) * | 2021-06-16 | 2023-08-04 | 成都唐源电气股份有限公司 | Extraction method of contact net wire abrasion area |
CN114926371B (en) * | 2022-06-27 | 2023-04-07 | 北京五八信息技术有限公司 | Vertical correction and vanishing point detection method and device for panorama and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20010048483A1 (en) * | 1995-09-08 | 2001-12-06 | Orad Hi-Tec Systems Limited | Method and apparatus for determining the position of a TV camera for use in a virtual studio |
US6400848B1 (en) * | 1999-03-30 | 2002-06-04 | Eastman Kodak Company | Method for modifying the perspective of a digital image |
US6873732B2 (en) * | 2001-07-09 | 2005-03-29 | Xerox Corporation | Method and apparatus for resolving perspective distortion in a document image and for calculating line sums in images |
US7046404B2 (en) | 1999-10-28 | 2006-05-16 | Hewlett-Packard Development Company, L.P. | Document imaging system |
JP3866600B2 (en) | 2002-03-27 | 2007-01-10 | 株式会社東芝 | Image processing apparatus and image processing method |
US20080260256A1 (en) * | 2006-11-29 | 2008-10-23 | Canon Kabushiki Kaisha | Method and apparatus for estimating vanish points from an image, computer program and storage medium thereof |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4010754B2 (en) * | 2000-08-10 | 2007-11-21 | 株式会社リコー | Image processing apparatus, image processing method, and computer-readable recording medium |
JP4145014B2 (en) * | 2001-01-11 | 2008-09-03 | 株式会社リコー | Image processing device |
JP4617993B2 (en) * | 2005-04-28 | 2011-01-26 | ソニー株式会社 | Image processing apparatus, image processing method, program, and recording medium |
JP2008077489A (en) * | 2006-09-22 | 2008-04-03 | Canon Inc | Image processor, method, program, and storage medium |
-
2007
- 2007-03-16 CN CN2007100883550A patent/CN101267493B/en not_active Expired - Fee Related
-
2008
- 2008-03-13 US US12/076,122 patent/US8170368B2/en not_active Expired - Fee Related
- 2008-03-14 JP JP2008066104A patent/JP4952625B2/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20010048483A1 (en) * | 1995-09-08 | 2001-12-06 | Orad Hi-Tec Systems Limited | Method and apparatus for determining the position of a TV camera for use in a virtual studio |
US6400848B1 (en) * | 1999-03-30 | 2002-06-04 | Eastman Kodak Company | Method for modifying the perspective of a digital image |
US7046404B2 (en) | 1999-10-28 | 2006-05-16 | Hewlett-Packard Development Company, L.P. | Document imaging system |
US6873732B2 (en) * | 2001-07-09 | 2005-03-29 | Xerox Corporation | Method and apparatus for resolving perspective distortion in a document image and for calculating line sums in images |
JP3866600B2 (en) | 2002-03-27 | 2007-01-10 | 株式会社東芝 | Image processing apparatus and image processing method |
US20080260256A1 (en) * | 2006-11-29 | 2008-10-23 | Canon Kabushiki Kaisha | Method and apparatus for estimating vanish points from an image, computer program and storage medium thereof |
Non-Patent Citations (10)
Title |
---|
Clark et al., "Rectifying perspective views of text in 3D scenes using vanishing points", Pattern Recognition, vol. 36 No. 11, pp. 2673-2686, 2003. * |
Lu et al., "Perspective rectification of document images using fuzzy set and morphological operations", Image and Vision Computing, vol. 23 No. 5, pp. 541-553, 2005. * |
Maurizio Pilu "Extraction of illusory linear clues in perspectively skewed documents", Hewlett-Packard Laboratories. |
Monnier et al., "Sequential Correction of Perspective Warp in Camera-based Documents", Proceedings of IEEE Conference on Document Analysis and Recognition, vol. 1, pp. 394-398, 2005. * |
Paul Clark et al. "Rectifying perspective views of text in 3D scenes using vanishing points", The Journal of the Pattern Recognition Society, Oct. 2003, pp. 2673-2686. |
Pilu, "Extraction of illusory linear clues in perspectively skewed documents", Proceedings of IEEE CVPR, pp. 363-368, 2001. * |
Shijian Lu et al. "Document Image Rectification Using Fuzzy Sets and Morphological Operators", 2004 International Conference on Image Processing (ICIP), pp. 2877-2880. |
Shijian Lu et al. "Perspective rectification of document images using fuzzy set and morphological operations" Image and Vision Computing vol. 23, 2005, pp. 541-553. |
Trinh et al., "Image-based Structural Analysis of Building using Line Segments and their Geometrical Vanishing Points", IEEE International Joint Conference SICE-ICASE, Feb. 26, 2007. * |
Yi-Chao Ma et al. "An Integrated Approach for Perspective Distortion Correction of Small-Square Document Images", PR & AI, vol. 19 No. 4, Aug. 2006, pp. 503-508. |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8885972B2 (en) * | 2008-04-03 | 2014-11-11 | Abbyy Development Llc | Straightening out distorted perspective on images |
US20120321216A1 (en) * | 2008-04-03 | 2012-12-20 | Abbyy Software Ltd. | Straightening Out Distorted Perspective on Images |
US20140307967A1 (en) * | 2008-04-03 | 2014-10-16 | Abbyy Development Llc | Straightening out distorted perspective on images |
US9477898B2 (en) * | 2008-04-03 | 2016-10-25 | Abbyy Development Llc | Straightening out distorted perspective on images |
US8285077B2 (en) * | 2008-07-15 | 2012-10-09 | Nuance Communications, Inc. | Automatic correction of digital image distortion |
US20100014782A1 (en) * | 2008-07-15 | 2010-01-21 | Nuance Communications, Inc. | Automatic Correction of Digital Image Distortion |
US20120219221A1 (en) * | 2011-02-25 | 2012-08-30 | Sony Corporation | System and method for effectively performing a scene rectification procedure |
US8379979B2 (en) * | 2011-02-25 | 2013-02-19 | Sony Corporation | System and method for effectively performing a scene rectification procedure |
US8897600B1 (en) * | 2013-12-20 | 2014-11-25 | I.R.I.S. | Method and system for determining vanishing point candidates for projective correction |
US20170208207A1 (en) * | 2016-01-20 | 2017-07-20 | Fujitsu Limited | Method and device for correcting document image captured by image pick-up device |
US10187546B2 (en) * | 2016-01-20 | 2019-01-22 | Fujitsu Limited | Method and device for correcting document image captured by image pick-up device |
US10140691B2 (en) | 2016-04-26 | 2018-11-27 | Abbyy Development Llc | Correcting perspective distortion in double-page spread images |
US10970845B2 (en) | 2017-06-14 | 2021-04-06 | Canon Kabushiki Kaisha | Image processing apparatus, image processing method, and storage medium |
Also Published As
Publication number | Publication date |
---|---|
JP4952625B2 (en) | 2012-06-13 |
CN101267493A (en) | 2008-09-17 |
US20080226171A1 (en) | 2008-09-18 |
CN101267493B (en) | 2011-01-19 |
JP2008257713A (en) | 2008-10-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8170368B2 (en) | Correcting device and method for perspective transformed document images | |
JP6000455B2 (en) | Form recognition method and form recognition apparatus | |
US8457403B2 (en) | Method of detecting and correcting digital images of books in the book spine area | |
US9430704B2 (en) | Image processing system with layout analysis and method of operation thereof | |
US8401333B2 (en) | Image processing method and apparatus for multi-resolution feature based image registration | |
US7835589B2 (en) | Photographic document imaging system | |
US7894689B2 (en) | Image stitching | |
US8611662B2 (en) | Text detection using multi-layer connected components with histograms | |
US9076056B2 (en) | Text detection in natural images | |
US8811751B1 (en) | Method and system for correcting projective distortions with elimination steps on multiple levels | |
US8897600B1 (en) | Method and system for determining vanishing point candidates for projective correction | |
JP2000105829A (en) | Method and device for face parts image detection | |
WO2009114967A1 (en) | Motion scan-based image processing method and device | |
CN108961262B (en) | Bar code positioning method in complex scene | |
US20130322758A1 (en) | Image processing apparatus, image processing method, and program | |
US8306335B2 (en) | Method of analyzing digital document images | |
KR101377910B1 (en) | Image processing method and image processing apparatus | |
US8913836B1 (en) | Method and system for correcting projective distortions using eigenpoints | |
JP6542230B2 (en) | Method and system for correcting projected distortion | |
US9094617B2 (en) | Methods and systems for real-time image-capture feedback | |
JP4859061B2 (en) | Image correction method, correction program, and image distortion correction apparatus | |
JP2011107878A (en) | Position detection apparatus and position detection method | |
Zayene et al. | Data, protocol and algorithms for performance evaluation of text detection in arabic news video | |
KR20170088370A (en) | Object recognition system and method considering camera distortion | |
JP4696239B2 (en) | Method and apparatus for correcting inclination of character string |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: FUJITSU LIMITED, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:YIN, XU-CHENG;SUN, JUN;FUJIMOTO, KATSUHITO;AND OTHERS;REEL/FRAME:020901/0350;SIGNING DATES FROM 20080326 TO 20080328 Owner name: FUJITSU LIMITED, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:YIN, XU-CHENG;SUN, JUN;FUJIMOTO, KATSUHITO;AND OTHERS;SIGNING DATES FROM 20080326 TO 20080328;REEL/FRAME:020901/0350 |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
FPAY | Fee payment |
Year of fee payment: 4 |
|
FEPP | Fee payment procedure |
Free format text: MAINTENANCE FEE REMINDER MAILED (ORIGINAL EVENT CODE: REM.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
LAPS | Lapse for failure to pay maintenance fees |
Free format text: PATENT EXPIRED FOR FAILURE TO PAY MAINTENANCE FEES (ORIGINAL EVENT CODE: EXP.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STCH | Information on status: patent discontinuation |
Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362 |
|
FP | Lapsed due to failure to pay maintenance fee |
Effective date: 20200501 |